INDEX

Explanations

disguise or impersonation

This neuron activates on words and word pieces that indicate someone is disguising themselves or impersonating another (e.g., masquerading, disguised, impersonators).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

-1.75

Or

-1.53

 táctil

-1.50

 dziś

-1.48

</em>

-1.48

hrten

-1.46

 dzisiaj

-1.44

ll

-1.38

And

-1.38

fielder

-1.36

POSITIVE LOGITS

 torba

1.52

 igre

1.48

ÁRIO

1.43

 diskusi

1.41

("/",

1.41

ógicos

1.39

1.36

 misi

1.35

Missi

1.35

 miliki

1.32

Activations Density 0.011%